Albino_BlacMan: How long of an exposure do you need at 0.005 lux. Isn't that an imperative piece of information?
I assume some of the better sensors out there can capture some kind of image at that level if you leave the shutter open long enough.
Reply to NezhinI am not sure on the requirements for how Gain is defined, some cameras do allow negative gains, but this might be like setting a camera to ISO 50 when the base ISO is 100. In practice it makes little difference as you will not be buying a camera with this sensor, unless it comes in your next car. If the same tech makes it into a still camera then the ISO value will be clear.
Reply to NezhinThe exposure was 1/60th at f/1.4. We only know the exposure for the 0.005 lux scene was with 72 dB of gain, the same scene shot at 1/60 f/1.4 at 0 dB gain had 400 lux of light and was normally exposed. What we do not know is what ISO level 0 or 72 dB corresponds to. Although given the light is an intensity ratio there should be about 12 EV difference between the two exposures. I assumed 0 dB to be ISO 100 but of course it could be a different number, 200 or 400 or even 3200.
Reply to Nezhin, not sure how you get 270,000,000. If you are using decibels gain as used for power ratios. 72dB would equal 15,848,932 times, set to log 2 you get 23.92 EV, with a base ISO of 100 that would make it ISO 1,6 billion. If you use intensity ratio then 72 db is 3981.1, log2 is 11.96 EV or ISO 400,000 with base ISO of 100. Of course my db calculation memories are bit fuzzy
Simple man: In the world of uneducated light engineers...... What exactly is a lux? And how does it compare to ISO?
Interesting design in my opinion. It's certainly small enough to be in a cellphone; should Sony decide to take it that route.
See http://en.wikipedia.org/wiki/Lux for a full explanation. In essences it is the amount of light illuminating an object.. Lux is a measure of the intensity of the light source (lumens) and the area over which it is spread.
ISO is the gain setting of the camera, with some adjustments to get it to meet the film standards. If you wish to use the same shutter speed and aperture to take an image you need to raise the ISO setting as the lux level falls. If you look at the example images in the story, Sony has an image shot at f/1.4 1/60th of a second exposure at 400 lux, at the cameras base ISO or 0db gain. They then reshot the image under 0.005 lux at 72db of gain. If the camera had a base ISO of 100 then the ISO used n the image would have been ISO400,000 but many video cameras use higher base ISO values, ISO 200, ISO 400 etc. which would increase the ISO level of the image. So it may have been ISO 800,000 or ISO 1,600,000
donCortizone: Very nice. If I were into EVF's I'd seriously consider this system. (Assuming they roll out more than another set of 35/50's.)
Zeiss said they are coming out in order of basic popularity. These are based on the ZM line, a check of B&H on that line, list them in popularity as:35mm f/250mm f/235mm f/2.8 (already covered by ZA lens21mm f/2.850mm f/1.515mm f/2.825mm f/2.828mm f/2.885mm f/421mm f/4.518mm f/4Pick 3 more after 50 and 35mm.
Wally626: Equivalence in general works, but tends to break down on the margins. Yes a 50mm f/1.4 lens in a m4/3 camera at ISO 100 can get close to the same image as a 100mm f/2.8 on a FF camera at ISO 400, even more similar if both are 16 MP sensors. So same number of pixels, same DoF, roughly same noise level at both pixel and total image levels, same perspective, same shutter speed. However, this assumes a perfect lens, the m4/3 lens has to have a better resolution than the FF lens, or both have to be better than their sensors. It also falls apart if you stick a f/1.4 lens on the FF, what m4/3 is equivalent? Or your have a Nikon 800 with 36 MP, etc.
It is also easier to make a sharper lens for a smaller image circle which helps. Phone lenses are in general very sharp in terms of lines per mm, but of course have very few mm to work with.
Equivalence in general works, but tends to break down on the margins. Yes a 50mm f/1.4 lens in a m4/3 camera at ISO 100 can get close to the same image as a 100mm f/2.8 on a FF camera at ISO 400, even more similar if both are 16 MP sensors. So same number of pixels, same DoF, roughly same noise level at both pixel and total image levels, same perspective, same shutter speed. However, this assumes a perfect lens, the m4/3 lens has to have a better resolution than the FF lens, or both have to be better than their sensors. It also falls apart if you stick a f/1.4 lens on the FF, what m4/3 is equivalent? Or your have a Nikon 800 with 36 MP, etc.
When a youth I went to Hawaii and me and a friend decide to body surf shore waves on Maui. When the big waves hit near the shore is was amazing, Standing in waist deep water, then bare sand, then crushed and spit up on the beach. Usually followed by the next wave whacking you in the head. Cudos for taking pictures in that type of surf.
Mike FL: FWIW:
Hopefully, this will start a new trend for all the manufactures to reduce the megapixels for sensors of all size for the sack of IQ.
The revolution of the A7s is not really in still images. It is a fine low light camera and will have very low per pixel noise compared to other cameras, but is not that much better or not better than the Nikon D4s and other FF bodies when the images are scaled to the same output. At low ISO the A7 or A7r would both be better as would many other cameras. Where the A7s is revolutionary is in the video. Being able to use all the pixels (at least from the 16:9 crop) gives the camera much more low light ability than any other camera available to consumers. It is not a difference of 10 or 20% it is several stops of difference.
Frank_BR: I do not know why so much importance is given to distortion and vignetting, which can be easily corrected by software, provided they are not excessive.
It also seems a bit amateurish that a laboratory as DXO Labs can not discount the effect of the camera sensor on the test of a lens. Lenses that are tested today will continue to be sold for several years, but the DXO Labs results will be obsolete tomorrow because the sensors are continuously being improved.
There is a MTF value for the sensor by itself but I do not know of anyone who publishes it or measures it directly. The easiest method would be to measure the lens separately then measure the lens sensor combination then divide by they lens value. The combined value is just the product of the two separate measurements. So if the lens is 0.8 and the sensor is 0.95 (at a given frequency) then the overall value would be 0.76. I think in practice the sensor is near 1 for most of the frequencies used. A 24 MP sensor on FF has 4000 vertical pixels, so 167 pixels per mm, or 83 pixel pairs per mm. Depending on how one approaches the issues of sampling and such as well as issues such as antialiasing filters as well as sensor leakage you should end up with pretty good but not perfect resolution with samples of 40 line pairs per mm, which is normally the sharpest tested. With 2X oversampling the 24 MP FF is good to 42 lp/mm. A 36 MP sensor would get you 51 lp/mm.
Beckler8: This is more like a variation than an upgrade model. Video gets a big upgrade but audio, with no hot shoe and no mic input gets a huge downgrade. (Don't bother mentioning external recorders.) As the review points out, some will like the lens variation. Others, like me, would much prefer more/longer zoom range at the expense of lens speed. Not everyone wants maximum low-light performance at the expense of being able to do nothing zoom-wise...yes, 2.9x zoom is essentially nothing, esp. in video where presumably none of the smart digital zoom is usable. I'd gladly trade the flash for the hot shoe too! There must be a way to have both though. Anyway on a camera of this stature then, there needs to be 2 (or more) versions, catering to different preferences.
There are 3 versions of this camera, plus the RX10 if you really need the better video plus hot shoe plus clickable or non-clickable aperture and of course the 24-200mm equivalent f/2.8 lens.
discbrake: Good job, Sony. But OM-D E-M1 over RX100 III for a walk-around camera. E-M1 is "small" for me. I'm more excited for the Alpha 7s that will release in July.
The A6000 with 16-50mm kit lens has roughly the same equivalent aperture in terms of total light, DoF effects and focal length as the RX100mIII. Cost the same amount of money, but the A6000 is larger and heavier, not as good video quality and no built in ND filters. So larger more versatile camera if you buy extra lenses versus small compact camera. Seems to be a fair trade to me. My wife would prefer the RX100mIII and I the A6000.
Thorgrem: How does this compare to the much cheaper Panasonic GH4?
4K requires external recorder. Quality will be higher than GH4 especially in low light, but will cost more and be a larger total package.
l_d_allan: I wonder if such a work-around would apply to the Sony A7 shutter-slap issue?
Just the A7r, the A7 has electronic first curtain shutter to eliminate the shutter shock issue.
Seems like the close shutter, open shutter, wait, use EFC might result in less vibration in images but it would seem to introduce more lag between shutter button push and the actual image being taken. Nice thing about EFC as used by Sony and I assume the other makers is that the delay between shutter press and image recording becomes very short.
The patent discusses some specific light placement, but in the end it is: get a white background, flood it with light, place object on semi-transparent table, take picture. The novelty may be in the specific light and baffle placement, but otherwise it has been done for a long time.
Victor Engel: Not 500 times longer than in NYC but 500 times longer than in the previous facility, which happened to be located in NYC.
What has cave paintings got to do with anything?
Mentioned in the video that they hope to preserve the images for 2000 years. So 1/500 would be 4 years. Some math errors there somewhere, I expect most prints would do good for 40 or 50 years in a typical home environment, so more like 20,000 years if 500 times.
JaimeA: Mr. Jeff Keller: Please make sure to mention the omission of the electronic level in your report/review of this important camera. The level is an essential tool. I still cannot understand why it has been removed.
Now that the camera has the standard hot shoe you can buy one of these hot shoe bubble levels. I will grant I have used the electronic level a few times on my Sony but not enough I would not buy the camera over it.
photophile: "3 separate built in LCD monitors for camera settings and metadata entry. Large 10 inch screen folds out for full resolution HD monitoring with wide viewing angle and high brightness."
Battery life ?
Looks like a pretty big battery hanging off the back.
Random Asian Guy: A lot of my photos are from low light situations, So this camera interests me quite a bit.
But I'm surprised it can only manage 5 fps for stills with only 12 megapixels.
Speculation is it using the A7 shutter, which may not be able to handle much more than 5 fps. I would think the electronics side could do it. Maybe too much shutter shake at high speeds, lot of moving mass in a small body to damp 10 times per second.